INDEX

Explanations

instances involving non-canonical narratives or alternate storylines

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_post_32k-oai

Configuration

jbloom/GPT2-Small-OAI-v5-32k-resid-post-SAEs/v5_32k_layer_8.pt

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

32,768

Data Type

torch.float32

Hook Name

blocks.8.hook_resid_post

Hook Layer

Architecture

standard

Context Size

Dataset

Skylion007/openwebtext

Activation Function

topk

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

mith

-0.74

ðŁĺ

-0.70

tsy

-0.67

NRS

-0.66

 phil

-0.64

 incent

-0.63

 volunt

-0.63

NPR

-0.63

IPM

-0.62

ertodd

-0.62

POSITIVE LOGITS

 dstg

0.84

å£«

0.73

isodes

0.72

 Series

0.72

atars

0.70

 Pikachu

0.68

Attributes

0.68

 Powered

0.68

 Characters

0.68

ãĥīãĥ©ãĤ´ãĥ³

0.68

Activations Density 0.825%

instances involving non-canonical narratives or alternate storylines

No Comments

No Known Activations